The vast majority of persons with substance use disorders go undetected and untreated. Proactive screening, brief intervention, and referral for treatment approaches (SBIRT) have tremendous potential for identifying and influencing undetected substance abuse, but efforts to incorporate them in the community have met with substantial obstacles related to training, time, costs, and acceptance. Two alternatives therefore merit serious consideration. First, computer-delivered approaches have shown clear effects on substance use outcomes (e.g., Hester, Squires, & Delaney, 2005; Ondersma, Svikis, & Schuster, 2007), and have tremendous advantages in terms of replicability, accessibility, and cost. Second-and most parsimoniously-assessment itself is increasingly demonstrating the ability to positively impact substance use (e.g., Clifford et al., 2007). Both approaches could obviate many of the above-noted obstacles, but have yet to be directly compared to traditional person-delivered SBIRT. Further, the significant effects of assessment on substance use outcomes means that there has yet to be a clear evaluation of the true effect of brief intervention approaches. The current proposal will therefore randomly assign 680 general medical urban outpatients into screening-only control (SC), assessment only (CA), assessment plus computer-delivered BI (CACI), and assessment plus therapist-delivered BI (CATI) conditions, and compare all groups at 1, 3, and 6 month follow-up on (a) substance use outcomes, as measured by self-report and urine toxicology, and (b) cost, using sophisticated economic analyses. Prior to this trial, thorough development research will use patient and expert panel feedback to modify existing and previously successful motivational software (developed by this group) to target primary care medical patients at risk for heavy/problem substance use. The proposed design would allow direct evaluation of two theoretically distinct approaches that, if similar in efficacy to the therapist- delivered condition, could greatly increase access to SBIRT. This design would also set the stage for the most stringent evaluation to-date of SBIRT's true efficacy in primary care. PUBLIC HEALTH RELEVANCE: To date, SBI interventions, while effective in research-based settings, have been too difficult to actually replicate in most medical settings. Hence most of the 22 million persons with drug problems in this country go undetected and untreated. The proposed study will compare two greatly simplified approaches to screening and brief intervention (both utilizing computers) to the traditional approach. If either of the two new approaches work even close to as well as the usual approach, their simplicity would mean that many more persons with drug or alcohol problems would get helped. [unreadable] [unreadable] [unreadable]